The role of Evidence Synthesis in COVID-19
Authors: Ciara Keenan, Damian Fogarty, Samantha Cheng and Chris Noone
Introduction: On 31 December 2019, the Wuhan Health Commission reported atypical pneumonia to the World Health Organisation (WHO) in patients who had been experiencing symptoms since mid-December. Shortly thereafter, a new RNA virus, with similarity to a coronavirus usually seen in bats, was confirmed as SARS-CoV-2. This novel virus causes a range of clinical problems from a mild flu-like illness to an overwhelming, often fatal, Severe Acute Respiratory Syndrome (SARS) from which its name derives. There is now a concerted research drive by hundreds, if not thousands, of teams worldwide to better understand the virus, its biology and the associated disease, now called COVID-19. Many are either developing or working towards the identification and production of effective treatments and, of course, a vaccine.
There is much uncertainty on how the pandemic will end. Some hope that all human-to-human transmissions are successfully interrupted and the virus will disappear. That was what effectively eradicated severe acute respiratory syndrome-related coronavirus (SARs-CoV) in 2003 (source). Others believe that COVID-19 might act like Middle East respiratory syndrome coronavirus (MERS-CoV), which has recurred sporadically since it was first reported in Saudi Arabia in 2012 (source). Or, some predict the worst outcomes, where it may take root in communities and countries and rapidly deplete populations for many years to come as was the case with the 1918 Spanish influenza (source).
Most experts agree that as the global spread of COVID-19 continues to grow, disease control will be challenging and this requires collaborative solutions and cooperative spirit from all groups.
The role of Evidence Synthesis specialists: Public concern has risen due to a combination of factors. Rumours and conjecture have spread through social and mainstream media, exacerbated by an inconsistent and non-standardised global healthcare response. Much of this is because scientific knowledge is incomplete but rapidly changing. This changing environment needs rapid knowledge synthesis and dissemination.
Evidence Synthesis specialists understand the importance of employing systematic and transparent methods to locate the available data in an unbiased way. In many ways, COVID-19 research is easy to find, with a proliferation of portals and repositories providing access to it and consistent terminology for the virus (SARS-CoV-2) and the disease (COVID-19).
However, difficulties have also been clear. First, the need for information has meant that publishers have relaxed some of their strict peer review guidelines. This has made quality appraisal more difficult and studies have been published without sufficient peer review and scrutiny. Second, the global impact of COVID-19 across all sectors of society (political, economic, personal liberties and freedom for example), means that research is being produced in large volumes across all disciplines, making it difficult to categorise.
An overview of Current Systematic Reviews and Meta-analyses: Teams of Evidence Synthesis experts have been working on various topics across the globe, as of 25 March 2020, at least 18 Systematic Reviews or Meta-Analyses were available (see them here). These reviews largely but not exclusively focus on medical aspects of the COVID-19 pandemic. They include the epidemiology and clinical characteristics, diagnostic features, potential treatments and outcomes for patients.
For example, Sun and colleagues (2020) provided a comprehensive picture of the clinical characteristics of COVID-19 patients by reviewing relevant studies and performing a single-arm meta-analyses. They estimated the incidence of fever (89% [0.818,0.945]), cough (72% [0.657,0.782]), muscle soreness or fatigue (43% [0.213,0.652]), acute respiratory distress syndrome (15% [0.046,0.296]), an abnormal chest CT (97% [0.921,0.993]), development of a critical condition (18% [0.127,0.243]) and death (4%; [0.027,0.061]).
Several potential risk factors which may aggravate the effects of COVID-19 have been investigated. Yang and colleagues (2020) conducted a systematic review and meta-analysis of the prevalence of comorbidities. They found that the most common comorbidities were hypertension (17% [14% – 22%]), diabetes (8% [6% – 11%]), cardiovascular diseases (5% [ 4% – 7%]) and respiratory system diseases (2% [1% – 3%]). Patients with severe symptoms of COVID-19 were more likely to have a co-morbidity of hypertension (Odds Ratio (OR) 2.36 [1.46-3.83])), respiratory system disease ,(OR 2.46 [1.76-3.44]) or cardiovascular disease (OR 3.42 [1.88-6.22]). A review of political and sociological studies of epidemics by Kapiriri and colleagues (2020) emphasised that disease outbreaks disproportionately affect vulnerable and marginalised communities. Notably, Ludvigsson (2020) reviewed 45 prevalence studies and found that children account for 1% to 5% of cases and that their experience of the disease is much milder than that of adults.
There have been few completed intervention studies of treatments for COVID-19. Some researchers have reviewed treatments for similar disease such as SARS and MERS (Zhang et al., 2020), and others have concluded that there are currently no pharmacological treatments for COVID-19 with high-level evidential support (Jiang, 2020).
The birth of a semi-automated evidence gathering bot: Dr Damian Fogarty is a senior physician and former researcher at Queen’s University Belfast and the UK Renal Registry. He is widely published and more recently been heavily involved with social media dissemination of medical and scientific information. Damian realised early on in the pandemic that the information around COVID-19 required rapid knowledge synthesis and dissemination. He contacted Ciara Keenan, an information retrieval specialist for Campbell UK and Ireland, who quickly built an automated aggregating twitter feed www.twitter.com/@COVID_Evidence.
This produces a diverse range of real-time, peer-reviewed/soon to be reviewed research and commissioned reports directly on a feed using the RSS sources from various science and medical databases. These include at present PubMed, F1000 research, BMC, bioRxiv, medRxiv, clinicaltrials.gov, Nature, Cell and Science. The search strategy includes terms such as: coronavirus, corona virus, 2019 coronavirus, corona virus disease, novel coronavirus, wuhan coronavirus, Coronavirus 2, COVID-19 and SARS-CoV2.
This database of emerging research includes a number of different resource types including:
- Case studies/case reports
- Guidance documents including clinical and normative practice guidance
- Reviews including narrative and systematic reviews
- Research reports including comparative studies, epidemiological studies, and translational research, as well as corrections and retractions
- Opinion pieces including editorials
Ciara noted early on that the research like the disease had a patter reflecting the spread of knowledge. With colleagues she developed an interactive geographical map reflecting emerging evidence sources (e.g. articles and resources) collated from sources such as the WHO database and the automated aggregating Twitter feed. The map geolocates where emerging research takes place to aid networking between research groups working across the world. This geolocation reflects the affiliation of researchers and not necessarily the geographic locations that are under study.
This map covers any COVID-19 related resources including all aspects of clinical research, public health, economics, social aspects, etc. It does not include news articles, nor resources about other coronaviruses, such as MERS and SARS. The interactive geographic map is powered by EviAtlas, an open access platform for visualizing synthesis data. As the situation is rapidly evolving and resources continue to emerge, this map and database is continually being updated.
Conclusion: After working on the database and map for more than a month, this work has been over-taken by other, much-better resourced teams. These include “living” databases of research from the EPPI-Centre, the Norwegian Institute of Public Health and Cochrane France. We have, therefore, decided to share what we have already done with others and to switch our efforts from trying to collate the vast amount of information being generated by others, to contributing to the evidence base itself. We are now working in collaboration with colleagues internationally on some of the systematic reviews that will be key to helping people make well-informed decisions during the COVID-19 pandemic and beyond.
Dr. Ciara Keenan is a research fellow at Queen’s University in Belfast with an established international reputation in evidence synthesis methodology, with a series of systematic review projects and expertise in the intersecting areas of health, social welfare, disability and education.
Dr Damian Fogarty is a senior physician and former researcher at Queen’s University and the UK Renal Registry. He is widely published and more recently been heavily involved with social media dissemination of medical and scientific information.
Dr. Samantha (Sam) Cheng is a biodiversity scientist at the Center for Biodiversity and Conservation at the American Museum of Natural History. She is an interdisciplinary conservation scientist whose work focuses on the link between nature and human health and well-being and builds tools and assessments for evidence-based conservation decisions. She runs Colandr (www.colandrapp.com), an open-access machine-learning assisted platform for evidence synthesis research efforts.
Dr. Chris Noone is a lecturer at the School of Psychology in NUI Galway. He has contributed to research on a range of topics broadly related to health and wellbeing. He is also a board member and chair of the research sub-committee for the National LGBT Federation in Ireland.